Grounded simulator-thinking (derived from best-minds): identifies who understands the topic best—or uses user-specified expert(s)—frames 3-5 facts to verify, runs targeted web search on named tech/products, then simulates that expert—not generic AI opinion. Uses 3-tier authenticity labels (verbatim quote / public…
A survey tool for taking over an existing or legacy software project. It reads tests, code, Git history, documentation, and spoken context to create a baseline of what is known, inferred, or still unverified.
Single user-facing entry / orchestrator for the proj- pipeline. Use when a request spans idea→ship, or the user doesn't know which step they're in, or wants a coordinated plan-execute-verify loop rather than calling one specialist directly. A thin Supervisor + bounded loop + facade over the fixed specialist skills…